CN109827777A - Rolling bearing fault prediction technique based on Partial Least Squares extreme learning machine - Google Patents
Rolling bearing fault prediction technique based on Partial Least Squares extreme learning machine Download PDFInfo
- Publication number
- CN109827777A CN109827777A CN201910259680.1A CN201910259680A CN109827777A CN 109827777 A CN109827777 A CN 109827777A CN 201910259680 A CN201910259680 A CN 201910259680A CN 109827777 A CN109827777 A CN 109827777A
- Authority
- CN
- China
- Prior art keywords
- rolling bearing
- signal
- squares
- partial
- learning machine
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The present invention relates to the rolling bearing fault prediction techniques of Partial Least Squares extreme learning machine.Multiple characteristic indexs such as time domain, frequency domain, time-frequency domain are analyzed first, propose the feature extracting method combined based on half normal distribution and experience wavelet de-noising, and fault diagnosis is carried out to rolling bearing and reaches better noise reduction effect close to original signal.Method of the improved ISOMAP nonlinear characteristic dimension reduction method of residual error in conjunction with fuzzy C-mean algorithm is proposed to rolling bearing fault Decline traits overall merit for more characteristic parameters again, improves rolling bearing in the variation tendency and nicety of grading of different decline stages.The data prediction model based on Partial Least Squares extreme learning machine is proposed based on extreme learning machine theory, is optimized parameter in ELM, is chosen optimal node in hidden layer and articulamentum weight, and choose Softmax activation primitive.In this way, precision of prediction is high, the calculating used time is short, and the characteristic value prediction effect after cluster is good.It is accurately predicted by the failure phase that above step carries out rolling bearing.
Description
Technical field
The invention belongs to the research fields of rotating machinery fault prediction technique, and in particular to be based on half normal distribution and EWT
The Noise reducing of data combined, and combine the technology of equidistant mapping to carry out dimension-reduction treatment to nonlinear properties using fuzzy C-mean algorithm
Method.
Background technique
Modern industrial equipment can exist since it is extensive, high complexity, multivariable, runs on the characteristics of closed-loop control
Lead to the not preventible influence factor of many of equipment fault.Therefore, a series of catastrophic failures caused by functions of the equipments are lost
It is not within minority.People also recognize that the state-detection of equipment and failure predication are reliably run in guarantee unit safety gradually, in advance
Play the role of in accident prevention vital.The rolling bearing components one of important as rotating machinery, operating status
Stability is directly related to the working performance of integral device.The abort situation of rolling bearing is effectively predicted, correctly identifies the axis of rolling
The performance degradation stage is held, is all of great significance to the health control of the incidence and equipment that reduce peril.
External aspect, in the 1960s, state monitoring of rolling bearing technology starts to rapidly develop.1962,
Gustafsson and Tallian research discovery can be by the peak change of bearing vibration acceleration signal come vibration bearing
Initial failure.1974, the D.R.Harting of Boeing Co. opened the pioneer in resonance demodulation technique field.1998
Year, Norden E.Huang et al. proposes a kind of novel signal processing method Hilbert-Huang transformation, by signal decomposition at
Multiple IMF components simultaneously carry out Hilbert transformation.
Domestic aspect, starts the correlative study for rolling bearing fault detection occur the 1980s.2014, Xu Yonggang
Deng the method for diagnosing faults for proposing to combine dual-tree complex wavelet transform and independent component analysis (ICA), this method is efficiently separated
Be extracted the characteristic information of rolling bearing combined failure.Sui Wentao etc. is proposed again based on EMD and maximum kurtosis deconvolution (MKD)
The method combined carries out maximum kurtosis deconvolution processing to reconstruction signal to enhance fault message.
The development foundation of modern machinery and equipment failure predication is the reason based on signal disposal and analysis and combination machine learning
It is realized by method.When mechanical equipment operation, since the influence of the factors such as site environment and operating condition variation will lead in signal
A large amount of noise can be adulterated, while it is non-stationary also to cause Dynamic Signal to occur.When the statistic of these non-stationary signals is
Varying function needs to handle signal by Time-Frequency Analysis Method, and Fourier analysis is confined to the signal linearly slowly converted；EMD
The problem of (empirical mode decomposition) is then modal overlap, and envelope is excessive or very few may all occur.EWT (experience wavelet transformation) is certainly
It adapts to divide the frequency spectrum of initial signal, while perpendicular band is constructed in respective demarcation interval by corresponding bandpass filter
Bandpass filter group, but there is a problem that deciding field obscured.
For the fault signature collection of vibration signal, there are information redundancies, and calculation amount is larger, in turn results in algorithm accuracy rate pole
Big the problem of reducing, it would be desirable to dimension be carried out to more characteristic parameters and about subtracted.PCA (principal component analysis) belongs to non_monitor algorithm,
Sample class information is not accounted for, each failure phase is mixed in together；KPCA (core principle component analysis) with kernel function by mutually tying
The space data projection to higher-dimension is closed, but the selection of kernel function seriously affects dimensionality reduction effect.ISOMAP (equidistant mapping) will
High dimensional feature data projection is into identical global low-dimensional coordinate, and after dimensionality reduction, though can obviously distinguish, there is also part aliasings
Phenomenon.And the method for using improved ISOMAP and fuzzy C-mean algorithm to combine has well the variation of rolling bearing decline stage
Clustering Effect, cluster it is intermediate relatively concentrate and without aliasing, it is higher to cluster intensity.
It is that mechanical equipment health repairs main realization rate first is that extract to the variation of the failure of mechanical equipment carry out prediction with
Assessment, is mainly had the prediction technique based on failure physical model, is commented using the life cycle load and failure mechanism knowledge of product
Estimate its reliability, can in some cases and it is non-required establish dynamic model, will so take time and effort；Prediction based on statistics
Method, information needed comes from various probability density functions, the accuracy that confidence interval can be predicted with objective description, but it is suitable
Regression equation type is sometimes difficult to find.Failure prediction method based on data-driven refers to through analysis input, output and shape
Relationship between state parameter obtains model, in the historical data the mapping relations of study input and output, and portion's building is non-in the inner
The linear nontransparent and non-model for special object, calculates future value, realizes failure predication.
To sum up, the method for first having to propose suitable signal de-noising rejects data noise information close to original signal.It connects
, it proposes suitable feature dimension reduction method, improves the nicety of grading of rolling bearing decline stage, and there are good non-destructive tests
Effect.Secondly, it is ensured that the accuracy of failure prediction algorithm preferably carries out the failure degradation trend of rolling bearing accurate pre-
It surveys.
Summary of the invention
Described in view of this, the present invention proposes the noise-reduction method combined based on half normal distribution and EWT (experience wavelet transformation),
Truly rolling bearing fault signal is diagnosed, has also been proposed the rolling bearing feature dimension reduction method based on C-ISOMAP,
The nicety of grading for improving rolling bearing decline stage, followed by the event based on PLS-ELM (Partial Least Squares extreme learning machine)
Hinder prediction model, precision of prediction is high, and the calculating used time is short, and the characteristic value prediction effect after cluster is good.
In order to achieve the above objectives, the invention provides the following technical scheme:
Rolling bearing fault prediction technique based on Partial Least Squares extreme learning machine, the method include the following steps:
Step 1；EWT signal de-noising based on half normal distribution sum；
Analysis conventional EWT carries out existing limitation when signal de-noising, proposes the theory that new EWT is combined with half normal distribution
Algorithm；
Step 2；Rolling bearing Data Dimensionality Reduction based on C-ISOMAP；
Rolling bearing fault Decline traits are evaluated for more characteristic parameters, and it is big to improve neighborhood to ISOMAP application residual error
It is small, it proposes method of the improved ISOMAP nonlinear characteristic dimension reduction method in conjunction with fuzzy C-mean algorithm, is entirely supervised in rolling bearing
During survey, rolling bearing Injured level is distinguished, rolling bearing degradation trend is predicted and classified, by rolling bearing
Normal vibration signal and ultimate failure fault-signal as training data, establish fuzzy C-means clustering model；
Step 3；Rolling bearing fault prediction based on PLS-ELM；
On the basis of extreme learning machine basic theories, the failure predication side based on Partial Least Squares extreme learning machine is proposed
Method optimizes the parameter in ELM using Partial Least Squares method, chooses optimal hidden layer neuron number and each articulamentum
Between weight.
The beneficial effect of the present invention compared with the existing technology is:
It is proposed the theoretical algorithm that half normal distribution is combined with EWT, original signal has obtained Accurate Reconstruction.Keeping EMD method
Completeness on the basis of, compared to EEMD screening the number of iterations reduce, improve signal-to-noise ratio more pasted to effectively inhibit noise
Nearly original signal, solves the problems, such as modal overlap.And improved ISOMAP nonlinear characteristic dimension reduction method is in conjunction with fuzzy C-mean algorithm
Method, more Precise Diagnosis goes out rolling bearing in the variation tendency of different decline stages.Compare after fuzzy C-means clustering
PCA, KPCA and improved ISOMAP effect picture and separability index it is found that improved have better Clustering Effect,
Cluster is intermediate relatively to be concentrated and without aliasing, and it is higher to cluster intensity.The fault prediction model data of PLS-ELM simultaneously
The degree of correlation is higher, and root-mean-square error is smaller, can be with each failure phase of Accurate Prediction by Wavelet Energy Spectrum entropy.
Detailed description of the invention
Fig. 1 is the route map of bearing vibration signal characteristic abstraction and failure predication research work.
The fault signature that Fig. 2 is EWT and half normal distribution combines extracts flow chart.
Fig. 3 is small echo signal decomposition figure.
Fig. 4 is CEEMD signal decomposition figure.
Fig. 5 is EWT signal decomposition figure.
Fig. 6 is half normal distribution-EWT noise reduction time-frequency domain figure.
Fig. 7 is the fuzzy C-mean algorithm flow chart of ISOMAP manifold learning.
Fig. 8 is the residual analysis figure of different faults degree.
Fig. 9 is different dimension reduction method comparison diagrams after fuzzy C-means clustering.
Figure 10 is the separability index comparison diagram of different dimension reduction methods.
Figure 11 is the regression machine prediction data based on Partial Least Squares extreme learning machine.
Figure 12 is the Wavelet Energy Spectrum entropy failure trend prediction figure based on PLS-ELM.
Specific embodiment
Specific embodiment 1: as shown in Figure 1, the rolling based on Partial Least Squares extreme learning machine of present embodiment
Bearing fault prediction technique, the method include the following steps:
Step 1；Signal de-noising based on half normal distribution and EWT；
Analysis conventional EWT carries out existing limitation when signal de-noising, proposes new half normal distribution (by a large amount of, independent uniform
The stochastic variable of small effects all obeys half normal distribution.Just because of this, mechanical breakdown field, mechanical oscillation are considered as partly just
State distribution.) theoretical algorithm that is combined with EWT；
Step 2；(fuzzy C-means clustering model is established, with ISOMAP epidemic algorithms to nonlinear data dimensionality reduction, place based on C-ISOMAP
Low-dimensional data after reason is able to maintain original topological relation) rolling bearing Data Dimensionality Reduction；
Rolling bearing fault Decline traits are evaluated for more characteristic parameters, and it is big to improve neighborhood to ISOMAP application residual error
It is small, propose that improved ISOMAP nonlinear characteristic dimension reduction method and fuzzy C-mean algorithm (to high dimensional feature data modeling, utilize person in servitude
Category degree evaluates a kind of iteration optimization clustering algorithm of sample point Clustering Effect superiority and inferiority) method that combines, it is entire in rolling bearing
In monitoring process, rolling bearing Injured level is distinguished, rolling bearing degradation trend is predicted and classified, by the axis of rolling
The normal vibration signal and ultimate failure fault-signal held establishes fuzzy C-means clustering model as training data；
Step 3；Rolling bearing fault prediction based on PLS-ELM；
On the basis of extreme learning machine basic theories, the failure predication side based on Partial Least Squares extreme learning machine is proposed
Method, using PLS, (Partial Least Squares is a kind of multipair Multilinear Regression modeling method, combines principal component analysis, canonical correlation
The reapective features of analysis and linear regression analysis, can provide more horn of plenty, deep information for prediction field) in ELM
Parameter optimizes, and chooses the weight between optimal hidden layer neuron number and each articulamentum.
Specific embodiment 2: present embodiment is the further explanation made to specific embodiment one；
Specific step is as follows for step 1 (signal de-noising based on EWT and half normal distribution):
Step 1 one；To the acceleration transducer setting sampling time being installed on rolling bearing pedestal and frequency.Then it determines and passes
Sensor channel number, and acquire the vibration signal under rolling bearing difference faulted condition.Then the fault-signal pre-processed,
Input signal as failure predication.
Step 1 two；For the problem that EWT deciding field obscures, the AM/FM amplitude modulation/frequency modulation component after division is more, lacks optimizing
Process needs to establish a screening index to select optimal component, optimizes here using half normal distribution, and expression formula is such as
Under:
In formula, σ represents the standard deviation of data.
Step 1 three；EWT combination half normal distribution is carried out to rolling bearing original signal to decompose.The threshold value obtained is to draw
Divide spectral boundaries, each frequency band establishes filter, and construction EWT de-noising signal obtains high frequency to the AM-FM component of low frequency, decomposites
To which the noise signal of rolling bearing usually contains in high-frequency signal, signal reconstruction, the signal after obtaining noise reduction are carried out.
Specific embodiment 3: present embodiment is the further explanation made to specific embodiment one；
Specific step is as follows for step 2 (the rolling bearing Data Dimensionality Reduction based on C-ISOMAP):
Step 2 one；In the entire monitoring process of rolling bearing, bearing Injured level can be not only distinguished, but also can be right
Rolling bearing degradation trend is predicted and is classified.Using the normal vibration signal of rolling bearing and ultimate failure fault-signal as
Training data establishes fuzzy C-means clustering model.
Step 2 two；Fault signature extracts: extracting initial damage stage, the moderate lesion stage, severe of data after noise reduction
Injury stage has dimension and nondimensional characteristic value, to form the huge feature set of representing fault signal.
Step 2 three；Take intrinsic epidemiological features: since the topological stability of ISOMAP is highly prone to the shadow of Size of Neighborhood k value
It rings, the residual error that application can retain structural information is determined.For the Injured level of rolling bearing, pass through after improving respectively
Realization huge feature set is mapped to lower dimensional space.Then rolling bearing performance assessment models are established: poly- by fuzzy C-mean algorithm
Class method finds out the cluster centre C={ C of fault-signal in the case of initial stage, moderate and three kinds of severe injury respectivelynormal,
Cfailure, then acquire degree of membership of each sample relative to normal sample.Dimensionality reduction effect passes through separability index JcEvaluation, expression
Formula is as follows:
In formula: tr represents the mark of matrix, SbFor inter _ class relationship matrix, SwWithin class scatter matrix.
Specific embodiment 4: present embodiment is the further explanation made to specific embodiment one；
Specific step is as follows for step 3 (the rolling bearing fault prediction based on PLS-ELM):
Step 3 one；On the basis of extreme learning machine basic theories, the event based on Partial Least Squares extreme learning machine is proposed
Hinder prediction technique, the parameter in ELM is optimized using Partial Least Squares method, choose optimal hidden layer neuron number and
Weight between each articulamentum.
Step 3 two；Prediction process based on Partial Least Squares extreme learning machine.Since conventional limit learning machine exists
Artificial parameter chooses the problem for causing precision of prediction low, first calculates winner with Partial Least Squares to the data after the dimensionality reduction of front
The weight w of component number r and each number of principal componentsi, the node in hidden layer of extreme learning machine is number of principal components r, each articulamentum power
It is again wi.In addition traditional activation primitive Sigmoid operation is complicated, and gradient is be easy to cause to disappear, and selects Softmax letter here
Number has better non-linear mapping capability, improves data fitting precision.
Step 3 three；Using the prediction model after optimization, unknown sample is predicted, obtains prediction result；
Embodiment 1:
In order to illustrate more clearly of the present embodiment, as shown in Figure 1, the rolling bearing event based on Partial Least Squares extreme learning machine
Hinder prediction technique, includes the following steps:
Step 1: Rolling Bearing Fault Character extracts and noise-reduction method research；
Step 2: the efficient feature dimensionality reduction of rolling bearing nonlinear data；
Step 3: the failure predication of rolling bearing Partial Least Squares extreme learning machine.
This example uses the open source data of U.S.'s Case Western Reserve University rolling bearing fault test platform, and equipment includes:
Motor, data logger, monitoring system, torque sensor and the power meter of 1.5KW, test fan end bearing are
SKF6203, motor speed 1797rpm, sample frequency 12KHz, bearing outer ring impaired loci is in 3 o'clock direction, fault diameter
For 0.5334mm.Housing washer Test to Failure data are collected by acceleration transducer, while being tried using electric spark
It tests on bearing and processes the Single Point of Faliure of equidirectional different lesion depths, respectively 0.178mm, 0.356mm, 0.533mm.To this
The fault type data of three kinds of Injured levels of same fault type, each 80 groups of samples of every kind of faulted condition, every group of sample
Contain 1024 data points again.The fault feature vector of three kinds of injury stages of rolling bearing, every group of each 7 feature are extracted respectively
Value, composition characteristic matrix are N=80 × 3 × 7=1680.
Firstly, the noise reduction of rolling bearing initial data:
Since the signal acquisition of rolling bearing is influenced by working environment (such as noise), cause original signal often non-linear,
Unstable.When covering useful information when noise signal is sufficiently large in rolling bearing, traditional EWT adopt when signal decomposition
It will cause the fuzzy problem of boundary with the boundary that maximum divides frequency band, thus cause the loss of high frequency useful signal, and therefore
Barrier vibration signal is to obey half normal distribution again.Therefore, using by the method for half normal distribution and EWT combination, pass through complementation
Mode solves the above problems, and Fig. 2 is characterized extraction flow chart.Comparison from Fig. 3 to Fig. 6 can be seen that new method more
Accurately be diagnosed to be the outer ring failure of rolling bearing, the processing of high-frequency signal is more perfect, guarantee signal noise filter out with it is complete
Whole property.
Secondly, the efficient feature dimensionality reduction of rolling bearing nonlinear data:
Basic procedure based on C-ISOMAP dimension-reduction algorithm is as shown in Figure 7.Since the topological stability of ISOMAP is big by neighborhood
The influence of small k value, Fig. 8 are the residual analysis figure in different faults stage, and as can be seen from the figure the residual error of minor failure is 9, in
Spending failure is 10, and severe failure is 6.Bearing difference will can be not only distinguished in rolling bearing during entire health monitoring
Degree of injury, and rolling bearing degradation trend can be predicted and be classified.By the normal vibration signal of rolling bearing and
Ultimate failure fault-signal establishes fuzzy C-means clustering model as training data.After feature vector normalized, to mostly special
Levy dimensionality reduction.J in Fig. 91For the separability index of minor failure, J2For the separability index of moderate failure, J3For severe failure can
Divide property index.As can be seen that using the method that improved ISOMAP and fuzzy C-mean algorithm combine to rolling bearing decline stage
Variation has a good Clustering Effect, cluster it is intermediate relatively concentrate and without aliasing, it is higher to cluster intensity.Figure 10 is that can divide
Property index comparison diagram, it can be seen that the dimensionality reduction effect of ISOMAP is more preferable in figure.
In addition, the rolling bearing fault prediction of Partial Least Squares extreme learning machine:
It uses lesion diameter for the housing washer fault data of 0.178mm, extracts the fault signature number of rolling bearing inner ring
According to 7, and preceding 3 main characteristic parameters obtained after improved ISOMAP dimensionality reduction and by fuzzy C-means clustering, composition
3 × 80=240 feature vector.The fault signature of test data set, can be preferably to the decline of rolling bearing as input
Cheng Jinhang is accurately predicted, obtains the regression machine prediction data figure of the Partial Least Squares extreme learning machine of Figure 11.Recycle small echo
The property of Energy Spectrum Entropy energy faults energy size obtains the prognostic chart in the rolling bearing different faults stage such as Figure 12.
Finally, it should be noted that the feasibility of examples detailed above only technical solution to illustrate the invention, although to the greatest extent may be used
Energy description is detailed, but on the basis of without departing from claims limited range, can still improve in certain details, hope the neck
Field technique personnel understanding.
Claims (4)
1. a kind of rolling bearing fault prediction technique based on Partial Least Squares extreme learning machine, which is characterized in that including with
Lower step:
Step 1；Signal de-noising based on half normal distribution-EWT；
Multiple characteristic indexs such as time domain, frequency domain, time-frequency domain are analyzed, the trouble diagnosibility of rolling bearing is reacted, propose new half
The theoretical algorithm that normal distribution is combined with experience wavelet transformation；
Step 2；Feature Dimension Reduction based on C-ISOMAP；
Rolling bearing fault Decline traits are evaluated for more characteristic parameters simultaneously, and to a variety of methods of Feature Dimension Reduction into
Row theory analysis proposes method of the improved ISOMAP nonlinear characteristic dimension reduction method of residual error in conjunction with fuzzy C-mean algorithm；It is rolling
In the entire monitoring process of dynamic bearing, bearing Injured level can be not only distinguished, but also can be to rolling bearing degradation trend
It is predicted and is classified；Using the normal vibration signal of rolling bearing and ultimate failure fault-signal as training data, mould is established
Paste C Clustering Model；
Step 3；Rolling bearing fault prediction based on PLS-ELM；
On the basis of extreme learning machine basic theories, the failure phase prediction based on Partial Least Squares extreme learning machine is proposed
Method optimizes the parameter in ELM using Partial Least Squares, chooses optimal node in hidden layer and each articulamentum power
Value, and choose Softmax activation primitive.
2. a kind of rolling bearing fault prediction side based on Partial Least Squares extreme learning machine according to claim 1
Method；Analysis bearing vibration signal is first had to directly to adopt when covering useful information when noise signal is sufficiently large in rolling bearing
Remove noise signal with experience wavelet de-noising, but experience wavelet transformation applies traditional maximum when dividing frequency band, it is tight to prop up
If support and width support adjacent frequency band, line of demarcation can fall in wide support, can not distinguish；It is characterized by: step 1 has
Steps are as follows for body:
Step 1 one；Since rolling bearing original vibration signal is the stochastic variable by a large amount of, independent uniformly small effects, clothes
From half normal distribution；Frequency band intervals number is determined using the formula of normal distribution:
In formula, σ represents the standard deviation of data；
Step 1 two；After determination section number, frequency band intervals are divided using EWT and obtain noise-containing high frequency AM-FM component,
And to its signal reconstruction；According to signal de-noising effect picture, determine that failure-frequency obtains abort situation.
3. it cannot reflect the variation in rolling bearing normal course of operation well due to single characteristic parameter, and multiple features
Often there is irrelevance and information redundancy in parameter, therefore using the method for Feature Dimension Reduction including all characteristic quantities
Space matrix carries out dimensionality reduction, using the rolling bearing feature dimension reduction method based on C-ISOMAP, it is characterised in that: step 2 is specific
Steps are as follows:
Step 2 one；Fault signature extracts: extracting the training sample initial damage stage, moderate loses stage, severe injury stage
Fault-signal in have dimension and nondimensional huge feature set；
Step 2 two；Intrinsic manifold feature extraction: the Size of Neighborhood of ISOMAP is improved, using residual error to improve the topology of ISOMAP
Stability；It is directed to the Injured level of rolling bearing simultaneously, is realized by improved ISOMAP reflect huge feature set respectively
It is mapped to lower dimensional space；
Step 2 three；Rolling bearing performance assessment models are established: finding out initial stage, moderate respectively by fuzzy C-means clustering method
Cluster centre C={ the C of fault-signal in the case of with three kinds of severe injurynormal,Cfailure, then acquire each sample relative to
The degree of membership of normal sample；
In order to accurate evaluation dimension reduction method, following dimensionality reduction Indexes of Evaluation Effect is used herein:
In formula, tr represents the mark of matrix, SbFor inter _ class relationship matrix, SwWithin class scatter matrix.
4. on the basis of extreme learning machine, being proposed based on Partial Least Squares extreme learning machine by two above step
Rolling bearing fault prediction technique optimizes the parameter in ELM using Partial Least Squares method, it is characterised in that: step
Three specific step is as follows:
Step 3 one；Training: the parameter in ELM is chosen by the number of principal components and its weight of Partial Least Squares, is looked for
Corresponding optimal node in hidden layer and articulamentum weight out, the training of complete paired data model；
Step 3 two；Test: according to the training pattern of acquisition, training dataset, the accuracy of detection building model are tested；
Step 3 three；Prediction: trained model is utilized, unknown sample is predicted, prediction result is obtained, using small echo
Energy Spectrum Entropy visualizes each failure phase.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910259680.1A CN109827777B (en) | 2019-04-01 | 2019-04-01 | Rolling bearing fault prediction method based on partial least square method extreme learning machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910259680.1A CN109827777B (en) | 2019-04-01 | 2019-04-01 | Rolling bearing fault prediction method based on partial least square method extreme learning machine |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109827777A true CN109827777A (en) | 2019-05-31 |
| CN109827777B CN109827777B (en) | 2020-12-18 |
Family
ID=66873979
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910259680.1A Expired - Fee Related CN109827777B (en) | 2019-04-01 | 2019-04-01 | Rolling bearing fault prediction method based on partial least square method extreme learning machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109827777B (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110160775A (en) * | 2019-06-04 | 2019-08-23 | 昆明理工大学 | A kind of high-pressure diaphragm pump one-way valve fault diagnostic method based on analysis of vibration signal |
| CN110398331A (en) * | 2019-07-22 | 2019-11-01 | 华侨大学 | Vibratory response frequency domain prediction method and device based on offset minimum binary |
| CN110672324A (en) * | 2019-09-02 | 2020-01-10 | 佛山科学技术学院 | Bearing fault diagnosis method and device based on supervised LLE algorithm |
| CN111553074A (en) * | 2020-04-27 | 2020-08-18 | 中国航空综合技术研究所 | Singular value elimination method for Weibull distribution fault sample |
| CN111652031A (en) * | 2019-12-02 | 2020-09-11 | 昆明理工大学 | Rolling bearing fault diagnosis method based on improved experience wavelet transform |
| CN111896254A (en) * | 2020-08-10 | 2020-11-06 | 山东大学 | Fault prediction system and method for variable-speed variable-load large rolling bearing |
| CN112116550A (en) * | 2019-06-20 | 2020-12-22 | 上海电机学院 | Insulator defect detection method based on multi-corner registration |
| CN112434558A (en) * | 2020-10-23 | 2021-03-02 | 北京航空航天大学 | Adaptive fault monitoring method based on +/-x sigma boundary interval model |
| CN112507479A (en) * | 2020-11-19 | 2021-03-16 | 电子科技大学 | Oil drilling machine health state assessment method based on manifold learning and softmax |
| CN113138080A (en) * | 2021-04-22 | 2021-07-20 | 东北大学 | Rolling bearing intelligent fault diagnosis method based on vibration twinning and extreme learning |
| CN113158769A (en) * | 2021-03-03 | 2021-07-23 | 安徽大学 | CEEMDAN and FastICA-based electromechanical device bearing vibration signal denoising method |
| CN114994109A (en) * | 2022-05-25 | 2022-09-02 | 电子科技大学 | XRF trace element quantitative analysis method based on ISOMAP-ELM |
| WO2023279382A1 (en) * | 2021-07-09 | 2023-01-12 | 徐萌萌 | Motor bearing running state fault detection method and system |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1811367A (en) * | 2006-03-03 | 2006-08-02 | 西安交通大学 | Integrated supporting vector machine mixed intelligent diagnosing method for mechanical fault |
| CN102393299A (en) * | 2011-08-02 | 2012-03-28 | 西安交通大学 | Method for quantitatively calculating operational reliability of rolling bearing |
| KR101667164B1 (en) * | 2015-10-02 | 2016-10-17 | 경북대학교 산학협력단 | Rotor speed-based bearing fault diagnosis method |
| CN106980761A (en) * | 2017-03-29 | 2017-07-25 | 电子科技大学 | A kind of rolling bearing running status degradation trend Forecasting Methodology |
| CN107036817A (en) * | 2017-04-05 | 2017-08-11 | 哈尔滨理工大学 | SVR rolling bearing performances decline Forecasting Methodology based on krill group's algorithm |
| KR20170093613A (en) * | 2016-02-05 | 2017-08-16 | 울산대학교 산학협력단 | Method for bearing fault diagnosis |
| CN107271187A (en) * | 2017-08-09 | 2017-10-20 | 西华大学 | A kind of method that quantitative Diagnosis is carried out to automobile speed variator bearing failure |
| CN108830218A (en) * | 2018-06-15 | 2018-11-16 | 哈尔滨工程大学 | A kind of underwater robot propeller method for diagnosing faults based on improvement Isomap algorithm ISOMAP |
| CN109444740A (en) * | 2018-11-14 | 2019-03-08 | 湖南大学 | A kind of the malfunction intellectual monitoring and diagnostic method of Wind turbines |
-
2019
- 2019-04-01 CN CN201910259680.1A patent/CN109827777B/en not_active Expired - Fee Related
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1811367A (en) * | 2006-03-03 | 2006-08-02 | 西安交通大学 | Integrated supporting vector machine mixed intelligent diagnosing method for mechanical fault |
| CN102393299A (en) * | 2011-08-02 | 2012-03-28 | 西安交通大学 | Method for quantitatively calculating operational reliability of rolling bearing |
| KR101667164B1 (en) * | 2015-10-02 | 2016-10-17 | 경북대학교 산학협력단 | Rotor speed-based bearing fault diagnosis method |
| KR20170093613A (en) * | 2016-02-05 | 2017-08-16 | 울산대학교 산학협력단 | Method for bearing fault diagnosis |
| CN106980761A (en) * | 2017-03-29 | 2017-07-25 | 电子科技大学 | A kind of rolling bearing running status degradation trend Forecasting Methodology |
| CN107036817A (en) * | 2017-04-05 | 2017-08-11 | 哈尔滨理工大学 | SVR rolling bearing performances decline Forecasting Methodology based on krill group's algorithm |
| CN107271187A (en) * | 2017-08-09 | 2017-10-20 | 西华大学 | A kind of method that quantitative Diagnosis is carried out to automobile speed variator bearing failure |
| CN108830218A (en) * | 2018-06-15 | 2018-11-16 | 哈尔滨工程大学 | A kind of underwater robot propeller method for diagnosing faults based on improvement Isomap algorithm ISOMAP |
| CN109444740A (en) * | 2018-11-14 | 2019-03-08 | 湖南大学 | A kind of the malfunction intellectual monitoring and diagnostic method of Wind turbines |
Non-Patent Citations (2)
| Title |
|---|
| SHAO HAIDONG，JIANG HONGKAI，LI XINGQIU，WU SHUAIPENG: "Intelligent fault diagnosis of rolling bearing using deep wavelet", 《KNOWLEDGE-BASED SYSTEMS》 * |
| 李瑞莲，兰媛，熊晓燕: "基于小波包变换和极限学习机的滚动轴承故障诊断", 《太原理工大学学报》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110160775A (en) * | 2019-06-04 | 2019-08-23 | 昆明理工大学 | A kind of high-pressure diaphragm pump one-way valve fault diagnostic method based on analysis of vibration signal |
| CN112116550B (en) * | 2019-06-20 | 2024-05-31 | 上海电机学院 | Insulator defect detection method based on multi-angle point registration |
| CN112116550A (en) * | 2019-06-20 | 2020-12-22 | 上海电机学院 | Insulator defect detection method based on multi-corner registration |
| CN110398331A (en) * | 2019-07-22 | 2019-11-01 | 华侨大学 | Vibratory response frequency domain prediction method and device based on offset minimum binary |
| CN110672324A (en) * | 2019-09-02 | 2020-01-10 | 佛山科学技术学院 | Bearing fault diagnosis method and device based on supervised LLE algorithm |
| CN110672324B (en) * | 2019-09-02 | 2021-03-26 | 佛山科学技术学院 | Bearing fault diagnosis method and device based on supervised LLE algorithm |
| CN111652031A (en) * | 2019-12-02 | 2020-09-11 | 昆明理工大学 | Rolling bearing fault diagnosis method based on improved experience wavelet transform |
| CN111553074A (en) * | 2020-04-27 | 2020-08-18 | 中国航空综合技术研究所 | Singular value elimination method for Weibull distribution fault sample |
| CN111896254A (en) * | 2020-08-10 | 2020-11-06 | 山东大学 | Fault prediction system and method for variable-speed variable-load large rolling bearing |
| CN112434558A (en) * | 2020-10-23 | 2021-03-02 | 北京航空航天大学 | Adaptive fault monitoring method based on +/-x sigma boundary interval model |
| CN112507479A (en) * | 2020-11-19 | 2021-03-16 | 电子科技大学 | Oil drilling machine health state assessment method based on manifold learning and softmax |
| CN112507479B (en) * | 2020-11-19 | 2022-03-15 | 电子科技大学 | Oil drilling machine health state assessment method based on manifold learning and softmax |
| CN113158769A (en) * | 2021-03-03 | 2021-07-23 | 安徽大学 | CEEMDAN and FastICA-based electromechanical device bearing vibration signal denoising method |
| CN113138080A (en) * | 2021-04-22 | 2021-07-20 | 东北大学 | Rolling bearing intelligent fault diagnosis method based on vibration twinning and extreme learning |
| WO2023279382A1 (en) * | 2021-07-09 | 2023-01-12 | 徐萌萌 | Motor bearing running state fault detection method and system |
| CN114994109A (en) * | 2022-05-25 | 2022-09-02 | 电子科技大学 | XRF trace element quantitative analysis method based on ISOMAP-ELM |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109827777B (en) | 2020-12-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109827777A (en) | Rolling bearing fault prediction technique based on Partial Least Squares extreme learning machine | |
| CN110849626B (en) | Self-adaptive sparse compression self-coding rolling bearing fault diagnosis system | |
| Cai et al. | Fault detection and diagnostic method of diesel engine by combining rule-based algorithm and BNs/BPNNs | |
| CN107036817B (en) | SVR rolling bearing performance decline prediction technique based on krill group's algorithm | |
| Xu et al. | A novel rolling bearing fault diagnosis method based on empirical wavelet transform and spectral trend | |
| CN101799368B (en) | Electromechanical device nonlinear failure prediction method | |
| Yongbo et al. | Review of local mean decomposition and its application in fault diagnosis of rotating machinery | |
| CN103245907B (en) | A kind of analog-circuit fault diagnosis method | |
| CN111504635B (en) | Planetary gear fault diagnosis method based on differential evolution probability neural network | |
| CN111238843B (en) | Fan health evaluation method based on rapid spectrum kurtosis analysis | |
| CN105424366A (en) | Bearing fault diagnosis method based on EEMD adaptive denoising | |
| Sheng et al. | Applications in bearing fault diagnosis of an improved Kurtogram algorithm based on flexible frequency slice wavelet transform filter bank | |
| CN112378660A (en) | Intelligent fault diagnosis method for aero-engine bearing based on data driving | |
| CN103558042A (en) | Rapid unit failure diagnosis method based on full state information | |
| Liu et al. | An online bearing fault diagnosis technique via improved demodulation spectrum analysis under variable speed conditions | |
| CN105626502A (en) | Plunger pump health assessment method based on wavelet packet and Laplacian Eigenmap | |
| Zhao et al. | Probabilistic principal component analysis assisted new optimal scale morphological top-hat filter for the fault diagnosis of rolling bearing | |
| Chen et al. | A visualized classification method via t-distributed stochastic neighbor embedding and various diagnostic parameters for planetary gearbox fault identification from raw mechanical data | |
| Xu et al. | An adaptive spectrum segmentation method to optimize empirical wavelet transform for rolling bearings fault diagnosis | |
| Zhao et al. | A novel deep fuzzy clustering neural network model and its application in rolling bearing fault recognition | |
| Li et al. | A Classification Algorithm of Fault Modes‐Integrated LSSVM and PSO with Parameters’ Optimization of VMD | |
| Zhao et al. | Research on constructing a degradation index and predicting the remaining useful life for rolling element bearings of complex equipment | |
| CN114263621A (en) | Test method and system for diagnosing and simulating cavitation fault of centrifugal pump | |
| Yang et al. | Transient fault diagnosis of track circuit based on MFCC-DTW | |
| Cui et al. | Egram based SVD method for gear fault diagnosis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20201218 Termination date: 20210401 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |